A nanoelectronics-blood-based diagnostic biomarker for ME/CFS (2019) Esfandyarpour, Davis et al

Reading the description of xCELLigence device, although it speaks of 'impedance' nothing in its description (that I can see anyway) suggests it is measuring anything more than simple resistance. i.e. No mention of any frequency-dependant reactance component, which the Stanford device definitely does measure. Impedance is the vector sum of a real resistance component and imaginary reactance component.

I start to get lost as soon as the description heads into the biology, but the Stanford description seems to be saying that one aspect of the biology influences resistance, versus a different aspect influencing reactance. I think from this they are suggesting they can sort the wheat from the chaff.

 

But if you had a system that was measuring thousands of batteries at a time, and could know statistically that approximately the same number of batteries would be connecting well enough for a reading, each time you take a reading, then you might get a degree of repeatability, even if not knowing how many batteries typically connect.

If you then do that for both healthy and unhealthy subjects, then you would not need too know the exact number of batteries, just the comparison based on assuming the number of batteries is statistically similar.

Along the way, if you gain enough confidence that the number of batteries is fairly similar each time, even though the actual number unknown, you would still have readings that could show if healthy or not.

We are talking battery voltages here, whereas the instruments are measuring impedances, but the principle at issue here is the same of course.

 

If you look at Photo (B) in Figure 2 it seems like it has a large amount of sensors. I'm not sure if it's hundreds or thousands of sensors. Also, on the same page it says that "the gold microelectrode biosensors in each well of ACEA’s electronic microtiter plates (E-Plates®) cover 70-80% of the surface area."

 

But surely that is not the point here. The point is that if you test using a tool with different dimensions or orientations or whatever there is no reason to think you will end up with a comparable result. If one system testing batteries has a millimetre between electrodes and another a metre the results will bear no relation to each other. Similarly if one tests batteries standing up and another lying down.

 

I've no way of knowing Jonathan, it was just a thought.

As regards orientations, that was why I was thinking that if enough measurements taken - 1000s - then there may be a good probability that a fairly similar number of cells, from measurement to measurement, might end up in a favourable orientation for being suitably oriented to make contact with the measurement probes. In the same way that if you keep throwing n x thousand of dice each time, a quite similar number will be six-up, similar five-up, etc, at each throw.

I don't understand what you mean re 1000:1 disparity of electrode distances? Silicon fabrication is vastly more repeatable than that I'm sure, though I don't know the numbers. Maybe I've got the wrong end of the stick on this.

 

It would not b an issue of reproducibility of a type of electrode but the possible variation between different products simply described as measuring 'impedance'. You can do that at any distance you like. If what matters is the impedance in a fluid layer adjacent to a short length of cell membrane of a particular configuration then you have to have exactly the right distance involved. Otherwise you might be measuring the impedance across a whole cell or across a membrane or in channels between pseudopodia or goodness knows what. You rightly point out that the nano needle may be measuring impedance in the true sense and the important component may be capacitance. That makes the micro anatomy very critical because the capacitance is likely to be due to the lipid bilayer of a piece of cell membrane. So a measurement across some fluid up against such a membrane might be very dependent on the integrity of the membrane.

The basic problem for me here is that I have no idea why impedance of any of these compartments should be aof any relevance to cell physiology or pathology.

 

Ah, now I see what you mean. Yes, you may well be right. But as a thought experiment, if you could measure a trillion cells at a time, would it matter? Would the randomisation mean things would be sufficiently statistically similar each time for the readings to be meaningfully consistent? And if so would it maybe be adequate with a few thousand?

And I would not know of course. Time will tell.

 

Yes, I'm sure that is true because some time back now I recall this being specifically mentioned, and the frequency response of something or other being discussed.

 

The nanoneedle paper is really hard to read and understand as relevant bits of the same subject are scattered all over the place. Tables and scatter plots would have made interpretation clearer.

Here are some snippets which indicate that Zre, the real in-phase component, is the part that changes the most, not Zimg the imaginary or capacitive part.

Key statement :

Average data measurements with the real resistive part bolded:

Overall range of data measurements with the real resistive part bolded:

|Z| is the total impedance with 15kHz, 250mV applied stimulus
Zre is the real or resistive portion of the impedance
Zimg is the imaginary or capacitive/inductive portion of the impedance

 

This is the dimension of the nanoneedle. The key sensing area in contact with the cell is 30nm, the separation between the two electrodes of the sensors.

Many people have built interdigitated electrodes in the cellular impedance analyser literature where the electrodes are side by side and typically 50-100um apart. This is much more manufacturable. That is ~1000x more separation between the electrodes than the nanoneedle, and roughly 10x larger than a PBMC cell size, and probably relies in the cells clumping together in a chain on top of the circuit board to create a measureable impedance.

Looking at pictures of various devices including the nanoneedle all electrodes of the sensors seem to be connected in parallel in one big circuit so that the many sensor areas are in fact chained together to form one sensor circuit.

Alain Moreau uses his CellKey cellular impedance analyser to group jurkat cells cells exposed to patient plasma by response to chemicals and presented data at the NIH conference and the Stanford Symposium in 2019. I have not come across any mention of him trying the salt test as in the nanoneedle and I am puzzled as to why.

 

There are also almost no paragraphs (generally a bad sign, as I think).

There is a paper on potassium currents in nerve cells of some brain areas. It says that nitric oxides changes the dynamics of voltage gated potassium channels, bringing more potassium out of the cell, thereby allowing the nerve cell to fire more often and therefore in a sustained manner. Steiner et al 2011

This could be related to the blood cells as currents within the cells might potentially behave different as well (because there is something in the blood, presumably a messenger molecule).

Funnily a friend of mine has been investigating potassium channels some 20 years ago, he explained to me: "There are chloride channels, which conduct into the cells, and a lot about them is known, but why there are potassium channels, which conduct outwards the cell, is a riddle." Unfortunately he didn´t find out, but it was found out only a bit later - according to wikipedia - that they serve as a counter regulation for the currents that go into the cell: sodium, chloride and calcium, with chloride being negative charged and therefore counteracting as well the sodium currents.

In my experience then, potassium and chloride are under some circumstances helpful, I need to check out this again and further. So, he might have missed to find the cure for my ME (this idiot).

 

Here are the app notes from the xCELLigence RTCA website showing different ways it can be used.

 

From memory Ron Davis published data showing that the impedence of all of the "controls" was lower than the "tests"i.e. cells + ME plasma. So at that level the results seem interesting; however, I think you would need to do a lot more tests and understand what is changing (i.e. in cells + ME plasma).

 

Thanks for the extra details @wigglethemouse.

A distance of 30nm really puzzles me. If I remember rightly that is the size of about five protein molecules.
And if it is resistance rather than capacitance that differs that too is strange. Unless one electrode is inside the cell and the other outside then it seems to me that what is being measured is the impedance of the solvent the cell is in, not the cell itself.

 

Interesting you should say that. I did some digging and found that the first Nanoneedle paper from 2012 was specifically for protein detection
Electrical Detection of Protein Biomarkers Using Nanoneedle Biosensors

Note : I didn't read the paywalled paper.

 

I've been trying to envisage how it would look in 3D with cells, plasma, and salt, in the channel.

So I went back and had a look at the 2013/2014 Nanoneedle Paper
Nanoelectronic Impedance Detection of Target Cells

In that paper the design of the nanoneedle geometry looks almost identical. Instead of gold electrodes they used 100nm thick polysilicon - future experiments must have shown gold is easier to depsoit, and has better conductivity.

Stackup :
TOP
20nm oxide layer
100nm electrode
30nm oxide isolation layer
100nm electrode
BOTTOM
---------------
250nm tall.

They have experiments where they show the reasoning for using 15kHz, the same used in the 2019 paper. They use water, salt solution, and yeast cells.

The picture on the right shows yeast cells clumped over the sensor. The sensor tip is under yeast cells in the area identified by the red dotted area. The dark shaded area is the micro-groove channel. Yeast cells are spread all over the place, not confined just to the micro-groove.


Which makes me wonder what the 3D view of cells and sensor tip looks like. Are cellular surfaces flexible enough to make good contact with both electrodes, 30nm apart. Would the cell flatten when it's in a dish or micro-groove channel such that it would have squarish sides that allows it to make good contact with both electrodes?

Aside : This confirms that the nanoneedle in the 2019 paper is not a specific design for ME patients. It was something they had lying around in the lab from previous work.

 

I am baffled to be honest.

 

And this is rather a very good sign, I think.

I think this happens often, that you don´t come up with a plan which then leads to a discovery, but you are aware of possibilities and even odd possibilities, and then by accident something almost stupid looking succeeds.

This actually may be the reason why the paper is written in a somehow confuse manner, if I remember rightly. They rather have no idea what it could mean, and stumbling around they try to hide their confusion. Although they may have been able to do some reasoning (and to do some better explanation on the technique maybe), it shows also that it isn´t completely by accident that Davis has a good Name (which hopefully helps along with the price he got any further).

 

[my bold]

I'm finding this very confusing @wigglethemouse. Wondering if it may be to do with these nanoneedle papers being a bit fast and loose with terminology?

Use of 'top' and 'bottom' with regards to the needle, I've been automatically thinking as meaning tip and base of the needle, but I realise that must be wrong, because a cell across the tip must be encountering both electrodes simultaneously for the last sentence to make sense. And Fig. 1b in the later paper also indicates the same. So the fabricated layers must be running lengthwise along each needle therefore. So the descriptions of 'bottom' and 'top' as if the needle is laying on its side when so described.

So if you were to look end-on (under a microscope obviously) at the tip of a needle, you would then see the layers arranged across it from one side to the other.

A 15kHz voltage signal is then injected across the two electrodes, and the needle on its own will exhibit its own impedance characteristics. If you then apply a test sample onto the tip, the impedance characteristics will then be modified by the sample.

Does this tally with your understanding?

But then: I'm still confused (dammit!). Because I don't see how you would fabricate what I just described? If you take a silicon slice then surely each layer is developed through the slice, not across it? In which case the needles would have to be manufactured laying across the slice, not sticking up from it, nor down into it (if they did that then the would be layered like the multi-fruit ice lollies, with different layers along their length). Very confused.

 

That is correct. My labelling of TOP and BOTTOM was in the Z direction.

This is the fabrication process described in the paper